@zhixuan92/multi-model-agent-core 3.9.1 → 3.10.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -4,7 +4,7 @@
4
4
  "naming": "claude-{tier}-{major}-{minor} | claude-{major}-{tier}-{date} (3.x legacy)",
5
5
  "rateSource": "https://docs.anthropic.com/en/docs/about-claude/models",
6
6
  "rateLookupDate": "2026-04-17",
7
- "defaults": { "supportsEffort": true, "inputTokenSoftLimit": 150000, "capabilities": ["web_search", "web_fetch"] },
7
+ "defaults": { "family": "claude", "supportsEffort": true, "inputTokenSoftLimit": 150000, "capabilities": ["web_search", "web_fetch"] },
8
8
  "profiles": [
9
9
  { "prefix": "claude", "tier": "standard", "cost": "medium", "input": 3, "output": 15, "bestFor": "general Claude tasks (catch-all, Sonnet rates)" },
10
10
  { "prefix": "claude-opus", "tier": "reasoning", "cost": "high", "input": 5, "output": 25, "bestFor": "high-ambiguity tasks where error cost is high" },
@@ -21,7 +21,7 @@
21
21
  "naming": "gpt-{major}.{minor}[-{size}] — sizes: pro, mini, nano (no suffix = base)",
22
22
  "rateSource": "https://developers.openai.com/api/docs/pricing",
23
23
  "rateLookupDate": "2026-04-25",
24
- "defaults": { "supportsEffort": true, "inputTokenSoftLimit": 128000, "capabilities": [] },
24
+ "defaults": { "family": "openai", "supportsEffort": true, "inputTokenSoftLimit": 128000, "capabilities": [] },
25
25
  "profiles": [
26
26
  { "prefix": "gpt", "tier": "standard", "cost": "medium", "input": 2.5, "output": 15, "bestFor": "general GPT family tasks (catch-all)" },
27
27
  { "prefix": "gpt-5", "tier": "reasoning", "bestFor": "reasoning-tier coding and agentic workflows", "inputTokenSoftLimit": 1000000, "capabilities": ["web_search"] },
@@ -40,7 +40,7 @@
40
40
  "naming": "gemini-{major}[.{minor}]-{tier}-{suffix} — tiers: pro ($2/$12), flash ($0.50/$3)",
41
41
  "rateSource": "https://cloud.google.com/vertex-ai/generative-ai/pricing",
42
42
  "rateLookupDate": "2026-04-17",
43
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 200000, "capabilities": ["web_search"] },
43
+ "defaults": { "family": "gemini", "supportsEffort": false, "inputTokenSoftLimit": 200000, "capabilities": ["web_search"] },
44
44
  "profiles": [
45
45
  { "prefix": "gemini", "tier": "standard", "cost": "low", "input": 0.5, "output": 3, "bestFor": "general Gemini tasks (catch-all, Flash rates)" },
46
46
  { "prefix": "gemini-3.1-pro", "tier": "reasoning", "cost": "medium", "input": 2, "output": 12, "bestFor": "long-context reasoning with large documents" },
@@ -53,7 +53,7 @@
53
53
  "naming": "grok-{major}[.{minor}]-{variant} — fast variant is ~15x cheaper than base",
54
54
  "rateSource": "https://docs.x.ai/developers/models",
55
55
  "rateLookupDate": "2026-04-17",
56
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 200000, "capabilities": ["web_search"] },
56
+ "defaults": { "family": "grok", "supportsEffort": false, "inputTokenSoftLimit": 200000, "capabilities": ["web_search"] },
57
57
  "profiles": [
58
58
  { "prefix": "grok", "tier": "standard", "cost": "low", "input": 0.2, "output": 0.5, "bestFor": "general Grok tasks (catch-all, fast-tier rates)" },
59
59
  { "prefix": "grok-4", "tier": "reasoning", "cost": "medium", "input": 3, "output": 15, "bestFor": "reasoning and analysis tasks" },
@@ -65,7 +65,7 @@
65
65
  "naming": "mistral-{tier}-{version} — tiers: large ($0.50/$1.50), small ($0.10/$0.30)",
66
66
  "rateSource": "https://mistral.ai/pricing",
67
67
  "rateLookupDate": "2026-04-17",
68
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
68
+ "defaults": { "family": "mistral", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
69
69
  "profiles": [
70
70
  { "prefix": "mistral", "tier": "standard", "cost": "low", "input": 0.5, "output": 1.5, "bestFor": "general Mistral tasks (catch-all, Large rates)" },
71
71
  { "prefix": "mistral-large", "bestFor": "cost-effective frontier-class coding" },
@@ -77,7 +77,7 @@
77
77
  "naming": "deepseek-{variant}[-{size}] — V4 sizes: flash, pro; legacy variants are flat-priced",
78
78
  "rateSource": "https://api-docs.deepseek.com/quick_start/pricing",
79
79
  "rateLookupDate": "2026-04-27",
80
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
80
+ "defaults": { "family": "deepseek", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
81
81
  "profiles": [
82
82
  { "prefix": "deepseek", "tier": "standard", "cost": "low", "input": 0.3, "output": 0.5, "bestFor": "cost-effective coding and reasoning" },
83
83
  { "prefix": "deepseek-v4-flash", "input": 0.14, "output": 0.28, "bestFor": "cheapest DeepSeek V4 with 1M context", "inputTokenSoftLimit": 1000000 },
@@ -89,7 +89,7 @@
89
89
  "naming": "llama-{major}-{variant} | meta-llama/llama-{major}-{variant} (OpenRouter)",
90
90
  "rateSource": "https://openrouter.ai/models",
91
91
  "rateLookupDate": "2026-04-17",
92
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
92
+ "defaults": { "family": "llama", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
93
93
  "profiles": [
94
94
  { "prefix": "llama", "tier": "standard", "cost": "low", "input": 0.15, "output": 0.6, "bestFor": "general open-weight model tasks" },
95
95
  { "prefix": "llama-4", "bestFor": "cost-effective open-weight coding" },
@@ -101,7 +101,7 @@
101
101
  "naming": "qwen{major}[.{minor}]-{tier} — no separator after 'qwen'",
102
102
  "rateSource": "https://openrouter.ai/qwen/qwen3.5-plus-02-15",
103
103
  "rateLookupDate": "2026-04-17",
104
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
104
+ "defaults": { "family": "qwen", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
105
105
  "profiles": [
106
106
  { "prefix": "qwen", "tier": "standard", "cost": "low", "input": 0.26, "output": 1.56, "bestFor": "cost-effective coding and multilingual tasks" }
107
107
  ]
@@ -111,7 +111,7 @@
111
111
  "naming": "glm-{major}[.{minor}] — GLM-5 is ~30% more expensive than GLM-4",
112
112
  "rateSource": "https://docs.z.ai/guides/overview/pricing",
113
113
  "rateLookupDate": "2026-04-17",
114
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
114
+ "defaults": { "family": "glm", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
115
115
  "profiles": [
116
116
  { "prefix": "glm", "tier": "standard", "cost": "medium", "input": 1, "output": 3.2, "bestFor": "general GLM/Zhipu family tasks" },
117
117
  { "prefix": "glm-5", "bestFor": "bilingual Chinese/English coding" },
@@ -123,7 +123,7 @@
123
123
  "naming": "kimi-k{generation}[.{minor}] | kimi-k{generation}-{variant}",
124
124
  "rateSource": "https://platform.kimi.ai/docs/pricing/chat",
125
125
  "rateLookupDate": "2026-04-17",
126
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
126
+ "defaults": { "family": "kimi", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
127
127
  "profiles": [
128
128
  { "prefix": "kimi", "tier": "standard", "cost": "low", "input": 0.6, "output": 3, "bestFor": "long-context tasks and coding" }
129
129
  ]
@@ -133,7 +133,7 @@
133
133
  "naming": "command-r[-plus] — plus variant is flagship",
134
134
  "rateSource": "https://cohere.com/pricing",
135
135
  "rateLookupDate": "2026-04-17",
136
- "defaults": { "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": ["web_search"] },
136
+ "defaults": { "family": "cohere", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": ["web_search"] },
137
137
  "profiles": [
138
138
  { "prefix": "command-r", "tier": "standard", "cost": "medium", "input": 2.5, "output": 10, "bestFor": "RAG workflows and grounded generation" }
139
139
  ]
@@ -143,9 +143,198 @@
143
143
  "naming": "MiniMax-M{major}.{minor}[-highspeed] — case-sensitive prefix",
144
144
  "rateSource": "https://platform.minimax.io/docs/guides/pricing-paygo",
145
145
  "rateLookupDate": "2026-04-17",
146
- "defaults": { "supportsEffort": true, "inputTokenSoftLimit": 200000, "capabilities": [] },
146
+ "defaults": { "family": "minimax", "supportsEffort": true, "inputTokenSoftLimit": 200000, "capabilities": [] },
147
147
  "profiles": [
148
148
  { "prefix": "MiniMax-M2", "tier": "standard", "cost": "low", "input": 0.3, "output": 1.2, "bestFor": "well-scoped coding and agent loops where cost matters", "avoidFor": "highest-stakes ambiguous work that needs top-tier judgment" }
149
149
  ]
150
+ },
151
+ {
152
+ "provider": "microsoft",
153
+ "naming": "phi-{major}[-{variant}]",
154
+ "rateSource": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/",
155
+ "rateLookupDate": "2026-04-29",
156
+ "defaults": { "family": "phi", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
157
+ "profiles": [
158
+ { "prefix": "phi", "tier": "standard", "cost": "low", "bestFor": "efficient open-weight Microsoft models" }
159
+ ]
160
+ },
161
+ {
162
+ "provider": "google-open",
163
+ "naming": "gemma-{major}[.{minor}] — open-weight Gemma family",
164
+ "rateSource": "https://ai.google.dev/gemma",
165
+ "rateLookupDate": "2026-04-29",
166
+ "defaults": { "family": "gemma", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
167
+ "profiles": [
168
+ { "prefix": "gemma", "tier": "standard", "cost": "low", "bestFor": "lightweight open-weight Google models" }
169
+ ]
170
+ },
171
+ {
172
+ "provider": "yi-01",
173
+ "naming": "yi-{tier}[-{version}]",
174
+ "rateLookupDate": "2026-04-29",
175
+ "defaults": { "family": "yi", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
176
+ "profiles": [
177
+ { "prefix": "yi", "tier": "standard", "cost": "low", "bestFor": "bilingual Chinese/English open-weight models" }
178
+ ]
179
+ },
180
+ {
181
+ "provider": "perplexity",
182
+ "naming": "sonar[-{variant}] — Perplexity's online-grounded models",
183
+ "rateSource": "https://docs.perplexity.ai/guides/pricing",
184
+ "rateLookupDate": "2026-04-29",
185
+ "defaults": { "family": "sonar", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": ["web_search"] },
186
+ "profiles": [
187
+ { "prefix": "sonar", "tier": "standard", "cost": "low", "bestFor": "online-grounded search-augmented tasks" }
188
+ ]
189
+ },
190
+ {
191
+ "provider": "amazon",
192
+ "naming": "nova[-{tier}][-{version}] — Amazon Bedrock native models",
193
+ "rateSource": "https://aws.amazon.com/bedrock/nova/pricing/",
194
+ "rateLookupDate": "2026-04-29",
195
+ "defaults": { "family": "nova", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
196
+ "profiles": [
197
+ { "prefix": "nova", "tier": "standard", "cost": "low", "bestFor": "Amazon Bedrock native multimodal models" }
198
+ ]
199
+ },
200
+ {
201
+ "provider": "ai21",
202
+ "naming": "jamba-{major}[.{minor}] — Mamba-transformer hybrid",
203
+ "rateLookupDate": "2026-04-29",
204
+ "defaults": { "family": "jamba", "supportsEffort": false, "inputTokenSoftLimit": 256000, "capabilities": [] },
205
+ "profiles": [
206
+ { "prefix": "jamba", "tier": "standard", "cost": "low", "bestFor": "long-context Mamba-transformer hybrid" }
207
+ ]
208
+ },
209
+ {
210
+ "provider": "ibm",
211
+ "naming": "granite-{major}[.{minor}]-{variant}",
212
+ "rateSource": "https://www.ibm.com/granite",
213
+ "rateLookupDate": "2026-04-29",
214
+ "defaults": { "family": "granite", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
215
+ "profiles": [
216
+ { "prefix": "granite", "tier": "standard", "cost": "low", "bestFor": "enterprise open-weight IBM models" }
217
+ ]
218
+ },
219
+ {
220
+ "provider": "nvidia",
221
+ "naming": "nemotron-{major}[-{variant}]",
222
+ "rateSource": "https://build.nvidia.com/explore/discover",
223
+ "rateLookupDate": "2026-04-29",
224
+ "defaults": { "family": "nemotron", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
225
+ "profiles": [
226
+ { "prefix": "nemotron", "tier": "standard", "cost": "low", "bestFor": "NVIDIA-optimized open-weight models" }
227
+ ]
228
+ },
229
+ {
230
+ "provider": "databricks",
231
+ "naming": "dbrx[-{variant}]",
232
+ "rateSource": "https://www.databricks.com/dbrx",
233
+ "rateLookupDate": "2026-04-29",
234
+ "defaults": { "family": "dbrx", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
235
+ "profiles": [
236
+ { "prefix": "dbrx", "tier": "standard", "cost": "low", "bestFor": "mixture-of-experts enterprise models" }
237
+ ]
238
+ },
239
+ {
240
+ "provider": "snowflake",
241
+ "naming": "arctic[-{variant}]",
242
+ "rateSource": "https://www.snowflake.com/en/blog/snowflake-arctic-open-efficient-foundation-language-models-snowflake/",
243
+ "rateLookupDate": "2026-04-29",
244
+ "defaults": { "family": "arctic", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
245
+ "profiles": [
246
+ { "prefix": "arctic", "tier": "standard", "cost": "low", "bestFor": "Dense-MoE hybrid Snowflake models" }
247
+ ]
248
+ },
249
+ {
250
+ "provider": "reka",
251
+ "naming": "reka-{tier}[-{version}]",
252
+ "rateLookupDate": "2026-04-29",
253
+ "defaults": { "family": "reka", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
254
+ "profiles": [
255
+ { "prefix": "reka", "tier": "standard", "cost": "low", "bestFor": "multimodal frontier models" }
256
+ ]
257
+ },
258
+ {
259
+ "provider": "ai2",
260
+ "naming": "olmo-{major}[.{minor}][-{variant}] — fully open (data + code + weights)",
261
+ "rateLookupDate": "2026-04-29",
262
+ "defaults": { "family": "olmo", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
263
+ "profiles": [
264
+ { "prefix": "olmo", "tier": "standard", "cost": "low", "bestFor": "fully open research models from AI2" }
265
+ ]
266
+ },
267
+ {
268
+ "provider": "nous-research",
269
+ "naming": "hermes-{major}[.{minor}][-{variant}]",
270
+ "rateLookupDate": "2026-04-29",
271
+ "defaults": { "family": "hermes", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
272
+ "profiles": [
273
+ { "prefix": "hermes", "tier": "standard", "cost": "low", "bestFor": "fine-tuned open-weight reasoning models" }
274
+ ]
275
+ },
276
+ {
277
+ "provider": "microsoft-wizardlm",
278
+ "naming": "wizardlm-{major}[.{minor}][-{variant}]",
279
+ "rateLookupDate": "2026-04-29",
280
+ "defaults": { "family": "wizardlm", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
281
+ "profiles": [
282
+ { "prefix": "wizardlm", "tier": "standard", "cost": "low", "bestFor": "instruction-following open-weight models" }
283
+ ]
284
+ },
285
+ {
286
+ "provider": "bigcode",
287
+ "naming": "starcoder{generation}[-{variant}]",
288
+ "rateSource": "https://huggingface.co/bigcode",
289
+ "rateLookupDate": "2026-04-29",
290
+ "defaults": { "family": "starcoder", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
291
+ "profiles": [
292
+ { "prefix": "starcoder", "tier": "standard", "cost": "low", "bestFor": "code-generation open-weight models" }
293
+ ]
294
+ },
295
+ {
296
+ "provider": "cognitive-computations",
297
+ "naming": "dolphin-{major}[.{minor}][-{variant}]",
298
+ "rateLookupDate": "2026-04-29",
299
+ "defaults": { "family": "dolphin", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
300
+ "profiles": [
301
+ { "prefix": "dolphin", "tier": "standard", "cost": "low", "bestFor": "uncensored open-weight models" }
302
+ ]
303
+ },
304
+ {
305
+ "provider": "openchat",
306
+ "naming": "openchat-{major}[.{minor}][-{variant}]",
307
+ "rateLookupDate": "2026-04-29",
308
+ "defaults": { "family": "openchat", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
309
+ "profiles": [
310
+ { "prefix": "openchat", "tier": "standard", "cost": "low", "bestFor": "open-source chat-tuned models" }
311
+ ]
312
+ },
313
+ {
314
+ "provider": "lmsys",
315
+ "naming": "vicuna-{major}[b][-{variant}]",
316
+ "rateLookupDate": "2026-04-29",
317
+ "defaults": { "family": "vicuna", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
318
+ "profiles": [
319
+ { "prefix": "vicuna", "tier": "standard", "cost": "low", "bestFor": "LLaMA-finetuned chat models from LMSYS" }
320
+ ]
321
+ },
322
+ {
323
+ "provider": "shanghai-ai-lab",
324
+ "naming": "internlm{major}[.{minor}][-{variant}]",
325
+ "rateLookupDate": "2026-04-29",
326
+ "defaults": { "family": "internlm", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
327
+ "profiles": [
328
+ { "prefix": "internlm", "tier": "standard", "cost": "low", "bestFor": "multilingual Chinese/English open-weight models" }
329
+ ]
330
+ },
331
+ {
332
+ "provider": "baichuan",
333
+ "naming": "baichuan{generation}[-{variant}]",
334
+ "rateLookupDate": "2026-04-29",
335
+ "defaults": { "family": "baichuan", "supportsEffort": false, "inputTokenSoftLimit": 128000, "capabilities": [] },
336
+ "profiles": [
337
+ { "prefix": "baichuan", "tier": "standard", "cost": "low", "bestFor": "Chinese-language open-weight models" }
338
+ ]
150
339
  }
151
340
  ]
@@ -214,9 +214,9 @@ export declare const VerifySkippedEvent: z.ZodObject<{
214
214
  event: z.ZodLiteral<"verify_skipped">;
215
215
  reason: z.ZodEnum<{
216
216
  no_command: "no_command";
217
+ other: "other";
217
218
  not_applicable: "not_applicable";
218
219
  dirty_worktree: "dirty_worktree";
219
- other: "other";
220
220
  }>;
221
221
  stage: z.ZodString;
222
222
  }, z.core.$strict>;
@@ -370,9 +370,9 @@ export declare const StageStatsMapSchema: z.ZodObject<{
370
370
  }>>;
371
371
  skipReason: z.ZodNullable<z.ZodEnum<{
372
372
  no_command: "no_command";
373
+ other: "other";
373
374
  not_applicable: "not_applicable";
374
375
  dirty_worktree: "dirty_worktree";
375
- other: "other";
376
376
  }>>;
377
377
  }, z.core.$strip>;
378
378
  spec_review: z.ZodObject<{
@@ -847,9 +847,9 @@ export declare const Event: z.ZodDiscriminatedUnion<[z.ZodObject<{
847
847
  event: z.ZodLiteral<"verify_skipped">;
848
848
  reason: z.ZodEnum<{
849
849
  no_command: "no_command";
850
+ other: "other";
850
851
  not_applicable: "not_applicable";
851
852
  dirty_worktree: "dirty_worktree";
852
- other: "other";
853
853
  }>;
854
854
  stage: z.ZodString;
855
855
  }, z.core.$strict>, z.ZodObject<{
@@ -11,8 +11,79 @@ import type { CostTier, ProviderConfig } from '../types.js';
11
11
  * Bare model names pass through unchanged.
12
12
  */
13
13
  export declare function extractCanonicalModelName(raw: string): string;
14
+ export declare const ModelFamilyEnum: z.ZodEnum<{
15
+ claude: "claude";
16
+ openai: "openai";
17
+ gemini: "gemini";
18
+ grok: "grok";
19
+ mistral: "mistral";
20
+ deepseek: "deepseek";
21
+ llama: "llama";
22
+ qwen: "qwen";
23
+ glm: "glm";
24
+ kimi: "kimi";
25
+ cohere: "cohere";
26
+ minimax: "minimax";
27
+ phi: "phi";
28
+ gemma: "gemma";
29
+ yi: "yi";
30
+ sonar: "sonar";
31
+ nova: "nova";
32
+ jamba: "jamba";
33
+ granite: "granite";
34
+ nemotron: "nemotron";
35
+ dbrx: "dbrx";
36
+ arctic: "arctic";
37
+ reka: "reka";
38
+ olmo: "olmo";
39
+ hermes: "hermes";
40
+ wizardlm: "wizardlm";
41
+ starcoder: "starcoder";
42
+ dolphin: "dolphin";
43
+ openchat: "openchat";
44
+ vicuna: "vicuna";
45
+ internlm: "internlm";
46
+ baichuan: "baichuan";
47
+ other: "other";
48
+ }>;
49
+ export type ModelFamily = z.infer<typeof ModelFamilyEnum>;
14
50
  export declare const modelProfileSchema: z.ZodObject<{
15
51
  prefix: z.ZodString;
52
+ family: z.ZodEnum<{
53
+ claude: "claude";
54
+ openai: "openai";
55
+ gemini: "gemini";
56
+ grok: "grok";
57
+ mistral: "mistral";
58
+ deepseek: "deepseek";
59
+ llama: "llama";
60
+ qwen: "qwen";
61
+ glm: "glm";
62
+ kimi: "kimi";
63
+ cohere: "cohere";
64
+ minimax: "minimax";
65
+ phi: "phi";
66
+ gemma: "gemma";
67
+ yi: "yi";
68
+ sonar: "sonar";
69
+ nova: "nova";
70
+ jamba: "jamba";
71
+ granite: "granite";
72
+ nemotron: "nemotron";
73
+ dbrx: "dbrx";
74
+ arctic: "arctic";
75
+ reka: "reka";
76
+ olmo: "olmo";
77
+ hermes: "hermes";
78
+ wizardlm: "wizardlm";
79
+ starcoder: "starcoder";
80
+ dolphin: "dolphin";
81
+ openchat: "openchat";
82
+ vicuna: "vicuna";
83
+ internlm: "internlm";
84
+ baichuan: "baichuan";
85
+ other: "other";
86
+ }>;
16
87
  tier: z.ZodEnum<{
17
88
  standard: "standard";
18
89
  reasoning: "reasoning";
@@ -30,6 +101,8 @@ export declare const modelProfileSchema: z.ZodObject<{
30
101
  supportsEffort: z.ZodBoolean;
31
102
  inputCostPerMTok: z.ZodOptional<z.ZodNumber>;
32
103
  outputCostPerMTok: z.ZodOptional<z.ZodNumber>;
104
+ cachedInputCostPerMTok: z.ZodOptional<z.ZodNumber>;
105
+ reasoningCostPerMTok: z.ZodOptional<z.ZodNumber>;
33
106
  rateSource: z.ZodOptional<z.ZodString>;
34
107
  rateLookupDate: z.ZodOptional<z.ZodString>;
35
108
  inputTokenSoftLimit: z.ZodNumber;
@@ -1 +1 @@
1
- {"version":3,"file":"model-profiles.d.ts","sourceRoot":"","sources":["../../src/routing/model-profiles.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,KAAK,EAAE,QAAQ,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAkB5D;;;;;;;;;GASG;AACH,wBAAgB,yBAAyB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAmB7D;AAKD,eAAO,MAAM,kBAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;iBAgB7B,CAAC;AAEH,MAAM,MAAM,YAAY,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,kBAAkB,CAAC,CAAC;AAkJ9D;;;;;GAKG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,MAAM,EAI1C,CAAC;AAEF,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,GAAG,YAAY,CAS9D;AAED,wBAAgB,qBAAqB,CAAC,OAAO,EAAE,MAAM,GAAG,CAAC,YAAY,GAAG,WAAW,CAAC,EAAE,CAErF;AAED,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,cAAc,GAAG,QAAQ,CAErE"}
1
+ {"version":3,"file":"model-profiles.d.ts","sourceRoot":"","sources":["../../src/routing/model-profiles.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,KAAK,EAAE,QAAQ,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AA2D5D;;;;;;;;;GASG;AACH,wBAAgB,yBAAyB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAY7D;AAID,eAAO,MAAM,eAAe;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAkCjB,CAAC;AACZ,MAAM,MAAM,WAAW,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,eAAe,CAAC,CAAC;AAE1D,eAAO,MAAM,kBAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;iBAmB7B,CAAC;AAEH,MAAM,MAAM,YAAY,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,kBAAkB,CAAC,CAAC;AA8J9D;;;;;GAKG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,MAAM,EAI1C,CAAC;AAEF,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,GAAG,YAAY,CAS9D;AAED,wBAAgB,qBAAqB,CAAC,OAAO,EAAE,MAAM,GAAG,CAAC,YAAY,GAAG,WAAW,CAAC,EAAE,CAErF;AAED,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,cAAc,GAAG,QAAQ,CAErE"}